sq 0.1.0 → 0.1.1

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: 12b002f2a938d9096d85360736c3949d8ca60935
-  data.tar.gz: 023f25cd005f28611644a9b6cecc89f703a50b61
+  metadata.gz: 8dc704917f0cdb8d045e5307a0962f84e5e4d8e8
+  data.tar.gz: 7cd4146a0f2216b369c463f871c842689f0cdcb5
 SHA512:
-  metadata.gz: 34ad6b35d57baa6b039ccca9a245f1675205b50a82317645414fafc91742c43c5a14ac230f36cdffe4103d8e06a6e452f04d77e19a54f3b1363a996e405f5d99
-  data.tar.gz: deda0c2bf3f2168db8ea07b52b348b310a4b6d810f4d1605ac41f3506f545648de6f21cc85882742ec2fcab273417c42486b082fb290ef6488e3c64bf7244566
+  metadata.gz: 595313fac368b5b7dd90dfd490e71b308f6e923086383e250b5d46b580cc52a20a02c1ebaf4dcb0f438f7582409fff7ea9a2e672ecd6ae03c78a72ed881701c6
+  data.tar.gz: 8f7d036a2b02df7931543eef834e2f340672b12905b3d4cefd555e2ac67d8448e7fc32464b77bc0cc77d92d2e256494ef4ab4fc829ca222167591f3c66c9a7be

data/bin/sq

@@ -16,6 +16,7 @@ EOS
 
   opt :directory, 'Choose the output directory', :short => '-o', :type => :string, :default => '.'
   opt :verbose, 'Print more info', :short => '-V', :type => :bool, :default => false
+  opt :format, 'Filename format', :short => '-F', :type => :string, :default => '%s.pdf'
 end
 
 if ARGV.empty?

data/lib/sq.rb

@@ -14,8 +14,9 @@ module SQ
       "SQ/#{version} +github.com/bfontaine/sq"
     end
 
-    # query an URI and return a list of PDFs. Each PDF is an hash with two
-    # keys: :uri is its absolute URI, :name is its name (last part of its URI).
+    # query an URI and return a list of PDFs. Each PDF is an hash with three
+    # keys: +:uri+ is its absolute URI, +:name+ is its name (last part of its
+    # URI), and +:text+ is each link text.
     # @uri [String]
     # @regex [Regexp]
     def query(uri, regex=/./)
@@ -24,23 +25,49 @@ module SQ
       doc = Nokogiri::HTML(open(uri, 'User-Agent' => user_agent))
       links = doc.css('a[href]')
 
-      uris = links.map { |a| URI.join(uri, a.attr('href')) }
-      uris.select! { |u| u.path =~ /\.pdf$/i && u.to_s =~ regex }
+      uris = links.map { |a| [a.text, URI.join(uri, a.attr('href'))] }
+      uris.select! { |_,u| u.path =~ /\.pdf$/i && u.to_s =~ regex }
 
-      uris.map do |u|
+      uris.map do |text,u|
         {
           :uri => u.to_s,
-          :name => u.path.split('/').last
+          :name => u.path.split('/').last,
+          :text => text
         }
       end
     end
 
+    # Output a formatted filename.
+    # @doc [Hash]   as returned from +SQ.query+.
+    # @fmt [String]
+    # @opts [Hash] additional info.
+    def format(doc, fmt='%s.pdf', opts={})
+      opts[:number] ||= 0
+      opts[:count]  ||= 0
+
+      fmt.gsub(/%./) do |f|
+        case f
+        when '%n' then opts[:number]
+        when '%N' then opts[:number]+1
+        when '%c' then opts[:count]
+        when '%s' then doc[:name].sub(/\.pdf$/i, '')
+        when '%S' then doc[:text]
+        when '%_' then doc[:text].gsub(/\s+/, '_')
+        when '%-' then doc[:text].gsub(/\s+/, '-')
+        when '%%' then '%'
+        end
+      end
+    end
+
     # query an URI and download all PDFs which match the regex. It returns the
     # number of downloaded PDFs.
     # @uri   [String]
     # @regex [Regexp] Regex to use to match PDF URIs
-    # @opts  [Hash]   Supported options: :verbose, :directory (specify the
-    #                 directory to use for output instead of the current one)
+    # @opts  [Hash]   Supported options: +:verbose+, +:directory+ (specify the
+    #                 directory to use for output instead of the current one),
+    #                 and +:format+ the output format. See the README for
+    #                 details.
+    #
     def process(uri, regex=/./, opts={})
       uris = self.query(uri, regex)
       count = uris.count
@@ -50,6 +77,7 @@ module SQ
       return 0 if uris.empty?
 
       out = File.expand_path(opts[:directory] || '.')
+      fmt = opts[:format] || '%s.pdf'
 
       unless Dir.exists?(out)
         puts "-> mkdir #{out}" if opts[:verbose]
@@ -57,12 +85,15 @@ module SQ
       end
 
       p = ProgressBar.create(:title => "PDFs", :total => count)
+      i = 0
 
       uris.each do |u|
-        open("#{out}/#{u[:name]}", 'wb') do |f|
+        name = format(u, fmt, {:number => i, :count => count})
+        i += 1
+        open("#{out}/#{name}", 'wb') do |f|
           open(u[:uri], 'rb') do |resp|
             f.write(resp.read)
-            p.log u[:name] if opts[:verbose]
+            p.log name if opts[:verbose]
             p.increment
           end
         end

data/lib/version.rb

@@ -3,7 +3,7 @@
 module SQ
   class << self
     def version
-      '0.1.0'
+      '0.1.1'
     end
   end
 end

data/tests/format_tests.rb

@@ -0,0 +1,63 @@
+# -*- coding: UTF-8 -*-
+
+class SQ_format_test < Test::Unit::TestCase
+
+  def setup
+    @foo = {
+      :text => 'Foo Bar',
+      :url  => 'http://example.com/foo.pdf',
+      :name => 'foo.pdf'
+    }
+
+    @opts = { :count => 42, :number => 0 }
+  end
+
+  def test_empty_format
+    assert_equal('', SQ.format(@foo, '', @opts))
+  end
+
+  def test_format_litteral
+    assert_equal('%', SQ.format(@foo, '%%', @opts))
+  end
+
+  def test_format_pdf_number0
+    assert_equal('0', SQ.format(@foo, '%n', @opts))
+  end
+
+  def test_format_pdf_number1
+    assert_equal('1', SQ.format(@foo, '%N', @opts))
+  end
+
+  def test_format_pdf_count
+    assert_equal('42', SQ.format(@foo, '%c', @opts))
+  end
+
+  def test_format_pdf_name
+    assert_equal('foo', SQ.format(@foo, '%s', @opts))
+  end
+
+  def test_format_link_text
+    assert_equal(@foo[:text], SQ.format(@foo, '%S', @opts))
+  end
+
+  def test_format_link_text_underscores
+    assert_equal('Foo_Bar', SQ.format(@foo, '%_', @opts))
+  end
+
+  def test_format_link_text_hyphens
+    assert_equal('Foo-Bar', SQ.format(@foo, '%-', @opts))
+  end
+
+  def test_format_no_special
+    assert_equal('foo-qux', SQ.format(@foo, 'foo-qux', @opts))
+  end
+
+  def test_format_multiple_percentsigns
+    assert_equal('%%%', SQ.format(@foo, '%%%%%%', @opts))
+  end
+
+  def test_format_multiple_placeholders
+    assert_equal('0-1-Foo-Bar', SQ.format(@foo, '%n-%N-%-', @opts))
+  end
+
+end

data/tests/query_tests.rb

@@ -27,8 +27,8 @@ class SQ_query_test < Test::Unit::TestCase
 
   def test_full_match
     pdfs = [
-      {:uri => "#{@http}/bar1.pdf", :name => 'bar1.pdf'},
-      {:uri => "#{@http}/bar2.pdf", :name => 'bar2.pdf'}
+      {:uri => "#{@http}/bar1.pdf", :name => 'bar1.pdf', :text => 'bar1'},
+      {:uri => "#{@http}/bar2.pdf", :name => 'bar2.pdf', :text => 'bar2'}
     ]
     assert_equal(pdfs, SQ.query("#{@url}/bar", /./))
     assert_equal(pdfs, SQ.query("#{@http}/bar", /./))
@@ -36,14 +36,14 @@ class SQ_query_test < Test::Unit::TestCase
 
   def test_absolute_path
     pdfs = [
-      {:uri => "#{@http}/bar1.pdf", :name => 'bar1.pdf'}
+      {:uri => "#{@http}/bar1.pdf", :name => 'bar1.pdf', :text => 'bar'}
     ]
     assert_equal(pdfs, SQ.query("#{@url}/ab/so/lu/te", /./))
   end
 
   def test_malformed_html
     pdfs = [
-      {:uri => "#{@http}/bar1.pdf", :name => 'bar1.pdf'}
+      {:uri => "#{@http}/bar1.pdf", :name => 'bar1.pdf', :text => 'bar'}
     ]
     assert_equal(pdfs, SQ.query("#{@url}/malformed1", /./))
     assert_equal(pdfs, SQ.query("#{@url}/malformed2", /./))

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: sq
 version: !ruby/object:Gem::Version
-  version: 0.1.0
+  version: 0.1.1
 platform: ruby
 authors:
 - Baptiste Fontaine
@@ -38,20 +38,6 @@ dependencies:
     - - ~>
       - !ruby/object:Gem::Version
         version: '2.0'
-- !ruby/object:Gem::Dependency
-  name: colored
-  requirement: !ruby/object:Gem::Requirement
-    requirements:
-    - - ~>
-      - !ruby/object:Gem::Version
-        version: '1.2'
-  type: :runtime
-  prerelease: false
-  version_requirements: !ruby/object:Gem::Requirement
-    requirements:
-    - - ~>
-      - !ruby/object:Gem::Version
-        version: '1.2'
 - !ruby/object:Gem::Dependency
   name: ruby-progressbar
   requirement: !ruby/object:Gem::Requirement
@@ -131,6 +117,7 @@ extra_rdoc_files: []
 files:
 - lib/sq.rb
 - lib/version.rb
+- tests/format_tests.rb
 - tests/process_tests.rb
 - tests/query_tests.rb
 - tests/tests.rb
@@ -160,6 +147,7 @@ signing_key:
 specification_version: 4
 summary: Bulk PDFs downloader
 test_files:
+- tests/format_tests.rb
 - tests/process_tests.rb
 - tests/query_tests.rb
 - tests/tests.rb